Google / Editorial analysis

Google connects Computational Discovery to agents

Google Research updated access to Computational Discovery, an experimental AI-supported tool for scientific research.

Published by Noosigma AI Original announcement Updated by Noosigma AI
Key points
  • The update lets approved testers use Computational Discovery through MCP in compatible agents.
  • The experimental tool combines ERA and AlphaEvolve to support scientific coding.
  • Access requires an application and approval; the examples come from Google Research.
Original publication: Google ↗
A crystalline blue and violet glass core connects translucent geometric modules in a bright setting.

Context and analysis by Noosigma AI

Editorial summary and context based on the cited source. Claims and results attributed to the organization are reports from that source; they do not constitute independent validation by Noosigma AI.

In an update published on November 3, the team said approved trusted testers can connect the system through MCP to compatible agents such as Antigravity and Claude Code. Access requires an application and approval.

Why it matters

Built with ERA and AlphaEvolve, Computational Discovery helps write and optimize scientific code. The post describes experimental applications in epidemiological forecasting, runoff estimates and atmospheric CO₂ mapping.

Google Research reported these examples and results. Access remains experimental and subject to approval, with no general availability announced. Source: Google Research, update of November 3, 2026.

Original publicationhttps://research.google/blog/empirical-research-assistance-era-from-nature-publication-to-catalyzing-computational-discovery/ ↗
← All news